🧠How We Think

When you search experience closely you cannot find a separate inner 'I' because the sense of a subject is just patterns of unobserved thoughts and attention habits that arise as if they were an owner of experience.

People often give reasons for their actions that aren't the true causes because System 1 produces behavior automatically and System 2 then constructs convincing post‑hoc explanations that feel like the cause.

When you switch between two speakers, attention doesn't flip instantly but briefly overlaps, because brain recordings show concurrent engagement with both the old and new speaker which lets you pick up the new voice before you lose the previous thread.

Higher cognitive ability can increase motivated reasoning because smarter people can build more sophisticated rationalizations that protect prior beliefs instead of correcting them.

Charismatic or warm people can make others trust them despite warnings because they elicit strong positive feelings in System 1 that can override skeptical, deliberative reasoning in System 2.

Constant internal chatter produces the feeling of a stable self because unobserved thoughts present themselves as 'me,' turning fleeting mental events into the illusion of an enduring actor.

Relying on reminders can weaken later unaided memory because removing the need to practice effortful encoding and retrieval prevents internal memory traces from forming, so people recall fewer tasks once reminders are gone.

Expert intuition is accurate only in domains where extensive practice and regular feedback let System 1 build reliable pattern recognition, which causes quick, skilled decisions in areas like chess but not in irregular real‑world settings.

The brain can follow more than one conversation in a noisy room because it partially splits attention and processes separate auditory streams at the same time, letting you extract multiple threads without fully focusing on any single speaker.

Joining elements that feel contradictory reveals design opportunities because refusing fixed categories lets surprising relationships form that boxed identities would hide.

Your sense of skill comes mainly from who surrounds you, because people use local, relative metrics like class rank and immediate peers to evaluate ability, so the same talent can feel exceptional in one group and mediocre in another — the 'frog pond' effect.

When policymakers impose abstract models on cities they change the territory to fit the map, which creates outcomes the model didn't account for and often produces unintended harms.

Cognitive illusions persist after correction because the feeling-based responses of System 1 remain compelling, so even when System 2 recognizes an error it can still 'feel' right.

Some people are better at following multiple conversations because they naturally sustain partial engagement with their current speaker while recruiting extra neural resources to sample others, enabling exploration without dropping focus.

Judging choices solely by outcomes misattributes decision quality because results create a retrospective bias that retrofits explanations to match the outcome, ignoring the original probabilities and process.

Seeing the self as an ongoing process rather than a fixed thing reduces attachment because noticing how 'selfing' shifts with context makes identities feel optional instead of absolute.

Using reminders creates a practical trade-off: they secure immediate completion but, because they skip the practice of retrieving intentions, they prevent the development of durable prospective-memory skills.

When a project assigns collective credit to everyone, people stop hoarding ideas because no single person can cash in on them, which causes freer sharing into a collaborative 'primordial soup' where the best concepts surface.

Tribalism creates echo chambers because groups define who is a trusted source, so disagreeing information is dismissed by source distrust rather than judged on its merits, blocking corrective facts.

Human thought uses two systems because many responses come up automatically from memory (System 1), while deliberate control and complex computations require slow, effortful processing (System 2).

Psychedelics can reveal that nondual or loving modes of awareness are possible, but because the drug state is transient those realizations must be mapped into everyday attention and practice to produce lasting change.

Framing beliefs as bets pushes people into probabilistic thinking because putting stakes on a claim forces you to assign probabilities, consider how you might be wrong, and seek information to refine your estimate.

During peak flow the mind stops narrating or anticipating, which removes the felt distance between experiencer and activity and makes the moment feel deeply rewarding.

Improvement can feel like destruction because entering a bigger, tougher environment often lowers your relative rank even as your absolute ability rises, which triggers loss, imposter feelings, and ego hits.

Decision uncertainty comes from two sources because missing information (information asymmetry) makes outcomes unpredictable and luck introduces random variation that can flip results even when decisions are sound.

Motivated reasoning causes people to defend priors because the drive to preserve existing beliefs makes them notice confirming evidence, expend effort to discredit contradictions, and slack on vetting supportive information.

Updating our maps is difficult because revising a model forces us to confront discrepancies between what we want to be true and what is true, which triggers psychological resistance.

Anecdotes can be more reliable than aggregate data when they disagree because firsthand experiences expose flaws in data collection or aggregation that can mask real problems.

People with stronger overlap in those brain signals can eavesdrop more effectively because concurrent engagement lets them monitor nearby conversations while staying involved in their own.

Meditation works on two levels: it can improve focus and reduce stress through repeated practice, or by turning attention onto attention it can show that a locateable, separate 'I' does not exist because inspection reveals only arising sensations and thoughts.

Effortful remembering strengthens memory because the extra mental demand of recalling without assistance triggers consolidation processes that embed information more deeply than passive or assisted recall.

Offloading one memory task does not free cognitive resources to improve other memory tasks because reminders do not reallocate attention or boost the internal encoding needed for tasks you still must remember, and they can even impair performance on the remaining tasks.

Covering a highway with a lid preserves traffic capacity while creating new developable land on top, which can be sold or built to help offset the high cost of repairing and upgrading the roadway.

The identity-shaking pain of being a beginner is actually a useful signal, because that discomfort is the cost of entry you pay when you put yourself into harder environments that raise standards and force real improvement.

High emotional arousal ('tilt') impairs reasoning because it suppresses frontal-lobe deliberative control, causing impulsive, lower-quality decisions until emotional intensity subsides.

Treating a model as the full truth causes overconfidence because it freezes policies around the abstraction and ignores changing realities and feedback loops, which reduces adaptability and leads to poor decisions.

Maps reflect their creators' biases and goals because mapmakers choose what to include, emphasize, or omit based on values and constraints, so models should be interpreted in the context of their origin.

When people are in cognitive ease, System 2 disengages, which leads them to act more impulsively, emotionally, and optimistically and to follow first impressions.

Stating ambitious targets out loud aligns priorities and normalizes overinvesting time or earnings, which mobilizes a team and raises the chance of landing the high-profile projects on that list.

Financial statements simplify because they aggregate thousands of transactions into a compact model, which omits qualitative signals like product quality and customer experience that can materially change a business's prospects.

Noticing a thought at the moment it appears breaks the automatic thought→emotion loop, so you can interrupt rumination and shorten how long an emotion like anger keeps running.

Reusing the same underlying idea in a different materialization creates distinct value because a new medium or technical approach manifests the composition in unique ways instead of merely duplicating it.

Staying in a comfortable, familiar setting protects your confidence but caps how far you go, because the small pond supplies validation without the tougher benchmarks needed to reveal and develop higher capacity.

Deconstruct decisions and record predicted outcomes before results because memorializing your decision tree and probability estimates prevents emotional retrofitting and lets you evaluate the decision process objectively.

Precommitments (Ulysses contracts) prevent predictable future biases because tying your future options to a prior plan reduces the power of later emotions and temptations to override long-term goals.

Maps are inherently imperfect because they deliberately omit details to create a usable reduction; if a map mirrored the territory perfectly it would cease to be a simplification and lose its purpose.

Deliberately slowing down engages System 2 because its effortful, analytical processing can catch mistakes that automatic System 1 responses miss, so pausing can prevent some—but not all—errors.

Use a small decision group with norms for accuracy, accountability, and diverse viewpoints because peers can spot each other's biases, prioritize truth-seeking over being right, and provide corrective feedback to improve models.

Chasing achievements delivers only fleeting satisfaction because wins are brief and new desires replace them, so anchoring well‑being in present-centered attention or the process itself gives more durable contentment.

Designing buildings doesn't just reflect life; it produces it, because the physical spaces architects create directly shape how people behave and live every day.

People interpret outcomes through reference-group lenses because protecting self-worth leads them to credit their own wins to skill while explaining others' wins as luck and others' losses as bad decisions, skewing judgment.

Apply universalism by mentally separating a claim from its messenger because judging the claim independently reduces source-based distrust and makes you evaluate the evidence on its own merits.

Even wide probabilistic ranges are useful because quantifying uncertainty (e.g., 20–80%) makes ignorance explicit and motivates information-seeking to narrow and improve predictions.

Externalizing intentions to reminders improves immediate task performance because offloading the need to keep future tasks in mind reduces mental effort and makes forgetting much less likely while the reminder exists.

Major architectural careers often peak late because expensive, high-risk projects are entrusted only to firms with long reputations, forcing designers to spend decades building the credibility and teams needed to compete.

Keeping a mission statement minimal preserves creative flexibility because a less reductive definition leaves wiggle room for unexpected directions and emergent work.

The field often pays poorly because it mixes artistic commitment with heavy technical liability, which pushes many practitioners to treat architecture as a vocation rather than a market-paid skill despite long training and big responsibility.

Models must be updated and remain flexible because their usefulness depends on matching reality—when reality shifts and the model doesn't, it stops predicting or explaining outcomes.

Physicists explicitly define where Newtonian or Einsteinian models apply because each model's assumptions only hold in certain domains, and delimiting those domains prevents using an inappropriate approximation that would break predictions.

Good maps are built through feedback because explorers' observations create correction loops that update the model to better match a changing territory.